Neonatal bactrial sepsis is a life-threatening disorder occuring in approximately 1-10 patients per 1000 live term births and almost three to four times more commonly in preterm infants (1). The mortality rate for neonatal bacterial sepsis varies between 20 and 75% depending on the organism, the immunocompetence of the host, and associated complications present at diagnosis (1,2). Nosocomial infections constitute a major cause of morbidity and mortality. Due to their immunologic immaturity, neonates are particularly vulnerable to the frequent invasive procedures and exosure to resistant organisms during their prolonged hospitalization. Reduced numbers of myeloid progenitor cells and circulating neutrophils contribute significantly to the developmental immaturity of neonatal host defense and to the increased susceptibility of the neonate to overwhelming infection. Future optimal therapy for prevention of neonatal nosocomial infection may include the use of adjuvant immunomodulator therapy. This trial will attempt to define the role of rhu GM-CSF as an immunomodulator of neonatal bone marrow myeloid progenitor activity, its influence on circulating peripheral neutrophil counts and their functional activity, and whether this results in the prevention of neonatal nosocomial infection. Any significant reduction in the incidence, morbidity and mortality of neonatal nosocomial infection will greatly reduce the number of hosopital days required for preterm newborns and reduce the amount of intensive care support associated with this disorder.